[unreadable] In the area of cancer, microarray analysis is being used to identify specific differences between normal and disease tissues. Such efforts are leading to the identification of small gene sets, or signatures, that can provide key information for classifying cancer types. We propose here a mechanism for translating these discoveries into cost-effective, quantitative assays based on the use of highly multiplexed, universal-primer-driven rtPCR (UP-rtPCR). In this proposal we will demonstrate and validate the combined use of artificial neural network analysis for identification of small gene signatures and UP-rtPCR technology for translating these signatures into high throughput assays for use in research and clinical settings. Success will be measured by the ability of this approach to differentiate 4 types of small, round, blue-cell tumors. This work will be performed in collaboration with Drs. Javed Khan and Gary Fogel. [unreadable] [unreadable] [unreadable]